Antipsychotic drugs and risks of myocardial infarction: a self-controlled case series study.

AIM: Antipsychotics increase the risk of stroke. Their effect on myocardial infarction remains uncertain because people prescribed and not prescribed antipsychotic drugs differ in their underlying vascular risk making between-person comparisons difficult to interpret. The aim of our study was to investigate this association using the self-controlled case series design that eliminates between-person confounding effects. METHODS AND RESULTS: All the patients with a first recorded myocardial infarction and prescription for an antipsychotic identified in the Clinical Practice Research Datalink linked to the Myocardial Ischaemia National Audit Project were selected for the self-controlled case series. The incidence ratio of myocardial infarction during risk periods following the initiation of antipsychotic use relative to unexposed periods was estimated within individuals. A classical case-control study was undertaken for comparative purposes comparing antipsychotic exposure among cases and matched controls. We identified 1546 exposed cases for the self-controlled case series and found evidence of an association during the first 30 days after the first prescription of an antipsychotic, for first-generation agents [incidence rate ratio (IRR) 2.82, 95% confidence interval (CI) 2.0-3.99] and second-generation agents (IRR: 2.5, 95% CI: 1.18-5.32). Similar results were found for the case-control study for new users of first- (OR: 3.19, 95% CI: 1.9-5.37) and second-generation agents (OR: 2.55, 95% CI: 0.93-7.01) within 30 days of their myocardial infarction. CONCLUSION: We found an increased risk of myocardial infarction in the period following the initiation of antipsychotics that was not attributable to differences between people prescribed and not prescribed antipsychotics.

Machine learning models in electronic health records can outperform conventional survival models for predicting patient mortality in coronary artery disease.

Prognostic modelling is important in clinical practice and epidemiology for patient management and research. Electronic health records (EHR) provide large quantities of data for such models, but conventional epidemiological approaches require significant researcher time to implement. Expert selection of variables, fine-tuning of variable transformations and interactions, and imputing missing values are time-consuming and could bias subsequent analysis, particularly given that missingness in EHR is both high, and may carry meaning. Using a cohort of 80,000 patients from the CALIBER programme, we compared traditional modelling and machine-learning approaches in EHR. First, we used Cox models and random survival forests with and without imputation on 27 expert-selected, preprocessed variables to predict all-cause mortality. We then used Cox models, random forests and elastic net regression on an extended dataset with 586 variables to build prognostic models and identify novel prognostic factors without prior expert input. We observed that data-driven models used on an extended dataset can outperform conventional models for prognosis, without data preprocessing or imputing missing values. An elastic net Cox regression based with 586 unimputed variables with continuous values discretised achieved a C-index of 0.801 (bootstrapped 95% CI 0.799 to 0.802), compared to 0.793 (0.791 to 0.794) for a traditional Cox model comprising 27 expert-selected variables with imputation for missing values. We also found that data-driven models allow identification of novel prognostic variables; that the absence of values for particular variables carries meaning, and can have significant implications for prognosis; and that variables often have a nonlinear association with mortality, which discretised Cox models and random forests can elucidate. This demonstrates that machine-learning approaches applied to raw EHR data can be used to build models for use in research and clinical practice, and identify novel predictive variables and their effects to inform future research.

Time spent at blood pressure target and the risk of death and cardiovascular diseases.

BACKGROUND: The time a patient spends with blood pressure at target level is an intuitive measure of successful BP management, but population studies on its effectiveness are as yet unavailable. METHOD: We identified a population-based cohort of 169,082 individuals with newly identified high blood pressure who were free of cardiovascular disease from January 1997 to March 2010. We used 1.64 million clinical blood pressure readings to calculate the TIme at TaRgEt (TITRE) based on current target blood pressure levels. RESULT: The median (Inter-quartile range) TITRE among all patients was 2.8 (0.3, 5.6) months per year, only 1077 (0.6%) patients had a TITRE ≥11 months. Compared to people with a 0% TITRE, patients with a TITRE of 3-5.9 months, and 6-8.9 months had 75% and 78% lower odds of the composite of cardiovascular death, myocardial infarction and stroke (adjusted odds ratios, 0.25 (95% confidence interval: 0.21, 0.31) and 0.22 (0.17, 0.27), respectively). These associations were consistent for heart failure and any cardiovascular disease and death (comparing a 3-5.9 month to 0% TITRE, 63% and 60% lower in odds, respectively), among people who did or did not have blood pressure 'controlled' on a single occasion during the first year of follow-up, and across groups defined by number of follow-up BP measure categories. CONCLUSION: Based on the current frequency of measurement of blood pressure this study suggests that few newly hypertensive patients sustained a complete, year-round on target blood pressure over time. The inverse associations between a higher TITRE and lower risk of incident cardiovascular diseases were independent of widely-used blood pressure 'control' indicators. Randomized trials are required to evaluate interventions to increase a person's time spent at blood pressure target.

Integrating Bio-ontologies and Controlled Clinical Terminologies: From Base Pairs to Bedside Phenotypes.

Electronic Health Records (EHR) are inherently complex and diverse and cannot be readily integrated and analyzed. Analogous to the Gene Ontology, controlled clinical terminologies were created to facilitate the standardization and integration of medical concepts and knowledge and enable their subsequent use for translational research, official statistics and medical billing. This chapter will introduce several of the main controlled clinical terminologies used to record diagnoses, surgical procedures, laboratory results and medications. The discovery of novel therapeutic agents and treatments for rare or common diseases increasingly requires the integration of genotypic and phenotypic knowledge across different biomedical data sources. Mechanisms that facilitate this linkage, such as the Human Phenotype Ontology, are also discussed.

The tip of the iceberg: challenges of accessing hospital electronic health record data for biological data mining.

Modern cohort studies include self-reported measures on disease, behavior and lifestyle, sensor-based observations from mobile phones and wearables, and rich -omics data. Follow-up is often achieved through electronic health record (EHR) linkages across primary and secondary healthcare providers. Historically however, researchers typically only get to see the tip of the iceberg: coded administrative data relating to healthcare claims which mainly record billable diagnoses and procedures. The rich data generated during the clinical pathway remain submerged and inaccessible. While some institutions and initiatives have made good progress in unlocking such deep phenotypic data within their institutional realms, access at scale still remains challenging. Here we outline and discuss the main technical and social challenges associated with accessing these data for data mining and hauling the entire iceberg.

Using big data from health records from four countries to evaluate chronic disease outcomes: a study in 114 364 survivors of myocardial infarction.

AIMS: To assess the international validity of using hospital record data to compare long-term outcomes in heart attack survivors. METHODS AND RESULTS: We used samples of national, ongoing, unselected record sources to assess three outcomes: cause death; a composite of myocardial infarction (MI), stroke, and all-cause death; and hospitalized bleeding. Patients aged 65 years and older entered the study 1 year following the most recent discharge for acute MI in 2002-11 [n = 54 841 (Sweden), 53 909 (USA), 4653 (England), and 961 (France)]. Across each of the four countries, we found consistent associations with 12 baseline prognostic factors and each of the three outcomes. In each country, we observed high 3-year crude cumulative risks of all-cause death (from 19.6% [England] to 30.2% [USA]); the composite of MI, stroke, or death [from 26.0% (France) to 36.2% (USA)]; and hospitalized bleeding [from 3.1% (France) to 5.3% (USA)]. After adjustments for baseline risk factors, risks were similar across all countries [relative risks (RRs) compared with Sweden not statistically significant], but higher in the USA for all-cause death [RR USA vs. Sweden, 1.14 (95% confidence interval 1.04-1.26)] and hospitalized bleeding [RR USA vs. Sweden, 1.54 (1.21-1.96)]. CONCLUSION: The validity of using hospital record data is supported by the consistency of estimates across four countries of a high adjusted risk of death, further MI, and stroke in the chronic phase after MI. The possibility that adjusted risks of mortality and bleeding are higher in the USA warrants further study.

Big biomedical data and cardiovascular disease research: opportunities and challenges.

Electronic health records (EHRs), data generated and collected during normal clinical care, are increasingly being linked and used for translational cardiovascular disease research. Electronic health record data can be structured (e.g. coded diagnoses) or unstructured (e.g. clinical notes) and increasingly encapsulate medical imaging, genomic and patient-generated information. Large-scale EHR linkages enable researchers to conduct high-resolution observational and interventional clinical research at an unprecedented scale. A significant amount of preparatory work and research, however, is required to identify, obtain, and transform raw EHR data into research-ready variables that can be statistically analysed. This study critically reviews the opportunities and challenges that EHR data present in the field of cardiovascular disease clinical research and provides a series of recommendations for advancing and facilitating EHR research.

Defining disease phenotypes using national linked electronic health records: a case study of atrial fibrillation.

BACKGROUND: National electronic health records (EHR) are increasingly used for research but identifying disease cases is challenging due to differences in information captured between sources (e.g. primary and secondary care). Our objective was to provide a transparent, reproducible model for integrating these data using atrial fibrillation (AF), a chronic condition diagnosed and managed in multiple ways in different healthcare settings, as a case study. METHODS: Potentially relevant codes for AF screening, diagnosis, and management were identified in four coding systems: Read (primary care diagnoses and procedures), British National Formulary (BNF; primary care prescriptions), ICD-10 (secondary care diagnoses) and OPCS-4 (secondary care procedures). From these we developed a phenotype algorithm via expert review and analysis of linked EHR data from 1998 to 2010 for a cohort of 2.14 million UK patients aged ≥ 30 years. The cohort was also used to evaluate the phenotype by examining associations between incident AF and known risk factors. RESULTS: The phenotype algorithm incorporated 286 codes: 201 Read, 63 BNF, 18 ICD-10, and four OPCS-4. Incident AF diagnoses were recorded for 72,793 patients, but only 39.6% (N = 28,795) were recorded in primary care and secondary care. An additional 7,468 potential cases were inferred from data on treatment and pre-existing conditions. The proportion of cases identified from each source differed by diagnosis age; inferred diagnoses contributed a greater proportion of younger cases (≤ 60 years), while older patients (≥ 80 years) were mainly diagnosed in SC. Associations of risk factors (hypertension, myocardial infarction, heart failure) with incident AF defined using different EHR sources were comparable in magnitude to those from traditional consented cohorts. CONCLUSIONS: A single EHR source is not sufficient to identify all patients, nor will it provide a representative sample. Combining multiple data sources and integrating information on treatment and comorbid conditions can substantially improve case identification.

Data resource profile: cardiovascular disease research using linked bespoke studies and electronic health records (CALIBER).

The goal of cardiovascular disease (CVD) research using linked bespoke studies and electronic health records (CALIBER) is to provide evidence to inform health care and public health policy for CVDs across different stages of translation, from discovery, through evaluation in trials to implementation, where linkages to electronic health records provide new scientific opportunities. The initial approach of the CALIBER programme is characterized as follows: (i) Linkages of multiple electronic heath record sources: examples include linkages between the longitudinal primary care data from the Clinical Practice Research Datalink, the national registry of acute coronary syndromes (Myocardial Ischaemia National Audit Project), hospitalization and procedure data from Hospital Episode Statistics and cause-specific mortality and social deprivation data from the Office of National Statistics. Current cohort analyses involve a million people in initially healthy populations and disease registries with ∼10(5) patients. (ii) Linkages of bespoke investigator-led cohort studies (e.g. UK Biobank) to registry data (e.g. Myocardial Ischaemia National Audit Project), providing new means of ascertaining, validating and phenotyping disease. (iii) A common data model in which routine electronic health record data are made research ready, and sharable, by defining and curating with meta-data >300 variables (categorical, continuous, event) on risk factors, CVDs and non-cardiovascular comorbidities. (iv) Transparency: all CALIBER studies have an analytic protocol registered in the public domain, and data are available (safe haven model) for use subject to approvals. For more information, e-mail s.denaxas@ucl.ac.uk.